Application And Research Of Chinese Medicine Cold Differentiation Diagnosis Based On Fuzzy-neural Network

Chinese medicine is a treasure of the Chinese culture,with a long story.and it is a theoretical and practical subject.Traditional Chinese Medicine(TCM) mainly use "four examination" and "eight principal" to perform disease diagnosis.However,TCM is difficult to learn and master because it has a strong subjectivity,and excessively relies on doctor’s clinical experience.So objectification of TCM was proposed,which try to use modern science and technology to develop some equipment used in TCM,such as Pulsations,which is used to analyze human pulse,tongue diagnosis instrument,which is used to analyze human tongue condition,and so on. With the rapid development of artificial intelligence(AI) and its penetration in the field of TCM recent years,giving an strong support to the objectification of TCM cold diagnosis,So it is necessary to further study for the objectification of TCM cold diagnosis.Cold is an exogenous diseases,which mainly resulted in exopathic wind or lays virus.clinical performance as cough,stuffy nose,running nose,chilliness with fever.In clinic it could be divided into six syndrome. Cold is a frequently encountered disease,which can occur in four reasons,but more frequent in winter and spring, light disease would recover without taking any medicine, however, it cannot be neglected because severe disease may kill one’s life. If we can combine TCM basic theories and AI and further the objectivity of TCM cold diagnosis, then good economy benefit may be produced. So it is worth the further study for the objectification of TCM cold diagnosis.This paper tries to implement the automatic diagnosis of cold syndrome by "four diagnosis method" based on fuzzy inference to make some contributions to the objectification of TCM cold syndrome diagnosis. Main symptom of cold selection and automatic diagnosis of cold will be discussed in this paper.Concerning about main symptom selection,there are a lot of symptoms for cold in clinic. If use all the symptom as the input of the ANFIS network,the structure of the network may become very large and complex. So it is necessary to pick out the main symptom from the symptom sets. In this paper, we try to get rid of those relatively minor symptom which has little or no effect on syndrome diagnosis,referring to the basic theory of TCM and combine the clinical illness case and the doctor’s clinical experience.Concering about automatic diagnosis of cold,a cold syndrome diagnosis method based on fuzzy inference is proposed. This method try to apply fuzzy inference system(FIS),artificial neural network(ANN) and related theory of TCM in the cold syndrome diagnosis in order to combine the advantage of FIS and ANN and the experimental knowledge of TCM. Its core is the use of Adaptive Neuro-Fuzzy Inference System(ANFIS).This method could be divided into the following five steps:Main symptom selection,determine the inference rules referring to the TCM theory,building FIS,converting the FIS to the related ANFIS,training the ANFIS with the sample data in clinic. At the end step,this paper use the BP algorithm to train the ANFIS network in order to optimizing the structure and the parameters of the network.